Active Appearance Model (AAM) based face tracking has advantages of accurate alignment, high efficiency, and effectiveness for handling face deformation. However, AAM suffers from the generalization problem and has difficulties in images with cluttered backgrounds. In this paper, we introduce two novel constraints into AAM fitting to address the above problems. We first introduce a temporal matching constraint in AAM fitting. In the proposed fitting scheme, the temporal matching enforces an inter-frame local appearance constraint between frames. The resulting model takes advantage of temporal matching's good generalizability, but does not suffer from the mismatched points. To make AAM more stable for cluttered backgrounds, we introduce a color-based face segmentation as a soft constraint. Both constraints effectively improve the AAM tracker's performance, as demonstrated with experiments on various challenging real-world videos.

In this work, we address the problem of tracking three dimensional (3D) faces and facial actions in video sequences. The main contributions of the paper are as follows. First, we develop an extended 2D+3D Active Appearance Models (AAM) based 3D face and facial action tracking framework using 2D view-based AAMs and a modified 3D face model. Second, we develop a robust shape initialization method based on local feature matching to track fast face motion. Experiments evaluating the effectiveness of the proposed algorithm are reported.

In this paper we present an approach for animating a virtual hand model with the animation data extracting from hand tracking results. We track the hand by using particle filters and deformable contour templates with a Web camera; then we extend the 2D tracking results into 3D animation data using inverse kinematics method; finally, local frame based method is proposed to simulate a 3D virtual hand with the 3D animation data. Our system performs in real time.

This paper proposes a method for fingerprint matching based on the local ridge similarity. Different from traditional methods which estimate rotation angle between a query and template minutia using orientation field, our approach estimates the local rotation angle directly from sampling points of ridges. After fitting samples by least square method, the optimal local rotation angle is obtained, moreover, the similarity between two ridges is measured by virtue of the fitting error. The best reference pair is then determined in the smoothed similarity histogram. Finally, a matching score is computed by projecting the query minutiae set to the template according to the best reference. The experimental results on the public fingerprint database, FVC2002 DB3(800 fingerprints, 8 impressions per finger) and a self-collected database(880 fingerprints, 5 impressions per finger), show that our approach is more effective compared with the conventional ridges-based approach.

Facial feature tracking is a crucial and challenging task in computer vision. Recently online-learning methods have become increasingly popular on account of their strong ability to adapt to variations and have achieved good results in tracking. However, all previous work used only raw intensity to build the model, which is very sensitive to condition changes. In this work, we present a real time, fully automatic facial feature detection and tracking approach using adaptive observation models based on edge structure, which is more reliable especially when the lighting state alters during tracking. Experimental results demonstrate that using edge map measures in observation modeling can improve the accuracy and robustness of tracking.

This paper proposes an automatic foreground segmentation system based on Gaussian mixture models and dynamic graph cut algorithm. An adaptive per-pixel background model is developed to set the data cost of an image graph. Shadow detection which is important to the foreground segmentation is introduced in this paper. A boundary smoothing algorithm is presented so as to eliminate edge aliasing and composite a foreground object onto a new background seamlessly. We verify our algorithm in different video sequences including some public video sequences. Experimental results demonstrate the efficiency of our proposed method.

This paper proposes a robust method to match distorted fingerprints. We first aligned the query and template fingerprints. To cope with the deformation brought by distorted fingerprints, we trained the distribution of orientation difference at a series of concentric circles around the corresponding pairs, and calculate the posterior probability of matched class using Bayesian method. We compute the score of unmatched absolute reliable minutiae by doubling the size of bounding box. Finally, we fuse the both scores with the score of matched minutiae using the multiplication rule. The experiments have been conducted on the public fingerprint collection, DB3 in FVC2002, and the results present that our method outperforms the conventional matching algorithm.

Gabor feature has been widely viewed as a good representation method for face recognition. AdaBoost is an excellent machine learning technique. Learning Gabor feature based classifier using AdaBoost is one of the best face recognition algorithms. However, dimensionality of Gabor feature space usually is very high, which makes the training program need huge memory or else take a very long time to run. In this paper, we propose a method which not only can solve the problem but also can improve recognition accuracy. Several subspaces with moderate size are randomly generated from original high dimensional Gabor feature space. Then strong classifier is trained in every random subspace (T. Kam Ho, 1998) respectively and the outputs of multiple classifiers are combined in the final decision. Experimental results demonstrate that the method saves a great amount of training time, and achieves an exciting recognition rate of 97.91% on the FERET Fb test set

We propose a novel face recognition algorithm based on dissociated region pair linear discriminant analysis (DRP-LDA). Given a training set, LDA is applied to every dissociated region pair of face images. The DRP-LDA features are obtained by projecting corresponding regions into subspace. Then classifier is constructed based on the features with appropriate distance measure. Multiple classifiers like this are combined together by weighted sum rule. Both regions and weights are learned by AdaBoost algorithm. Experimental results on FERET face databases show an impressive accuracy of our algorithm compared with conventional and component-based LDA methods.

In this paper, we propose a novel learning method, called Jensen-Shannon Boosting (JSBoost) and demonstrate its application to object recognition. JSBoost incorporates Jensen-Shannon (JS) divergence [Y. Bubner et al. (2001)] into AdaBoost learning. JS divergence is advantageous in that it provides more appropriate measure of dissimilarity between two classes and it is numerically more stable than other measures such as Kullback-Leibler (KL) divergence (see [Y. Bubner et al. (2001)]). The best features are iteratively learned by maximizing the projected JS divergence, based on which best weak classifiers are derived. The weak classifiers are combined into a strong one by minimizing the recognition error. JSBoost learning is demonstrated with face object recognition using a local binary pattern (LBP) [M. Pietikainen et al. (2004)] based representation. JSBoost selects the best LBP features from thousands of candidate features and constructs a strong classifier based on the selected features. JSBoost empirically produces better face recognition results than other AdaBoost variants such as RealBoost [R.E. Schapire et al. (1998)], GentleBoost [J. Friedman et al. (2000)] and KL-Boost [C. Liu et al. (2003)], as demonstrated by experiments.

Mark Weiser described a hypothetical world in which humans and computers were seamlessly united in a seminal article a decade ago; he called this vision ubiquitous computing (which now is also called pervasive computing). After that, technology has advanced along many dimensions, especially in hardware progress and wireless communication technologies. A number of leading technological organisations are exploring pervasive computing. But it is far from Weiser's vision become reality. Pervasive computing will be the future. Pervasive computing will be a fertile source of challenging research problems in computer systems for many years to come. But until today, there is less attention to the system evolution in pervasive computing environment. This paper discusses the current development of pervasive computing and technical challenges in deploying pervasive computing systems. First, a model for evolution of pervasive Information technology infrastructure of pervasive information is presented which shows how the existing technologies solutions available can be used to develop a pervasive information infrastructure. Then a generic architecture of pervasive computing system is given. Based these, an evolutionary system design approach is suggested.

In this paper, a novel method is introduced to estimate height from surface normal and implemented in a photometric 3D reconstruction system, base on the improved optimization function in the height domain. To avoid the global distortion from local integration, a pyramid reconstruction ideal is proposed, which sample the surface every certain points to form the separated point sets called sub-surfaces, and then step by step the coarser sub-surfaces is combined into finer ones and eventually derive the whole surface. Experiment shows that our method is faster and more accurate than some of the main method in this field.

Accurate localization of representative points of a face is crucial to many face analysis and synthesis problems. Active shape model (ASM) is a powerful statistical tool for face alignment. However, it suffers from variations of pose, illumination and expressions. In this paper, we analyze the mechanism of active shape model and realize that the ability of normal profiles to describe the local appearance pattern is very limited. For efficient appearance pattern representation, the local binary pattern is used and extended to describe the local patterns of facial key points. For the purpose of retaining the spatial information, sub-images of key points are divided into several regions, which are combined to define the extended local binary pattern (ELBP) histogram. Then we propose an improved ASM method framework, ELBP-ASM, in which local appearance patterns of key points are modelled using extended local binary pattern. Experimental results demonstrate that ELBP-ASM achieves more accurate results compared with original method used in ASM.

Keyboards, mice, and joy sticks are the most popular controlling and navigation devices in current 3D game. However, they are quite unnatural. In this paper, we propose a novel scheme to estimate the user's head pose and the estimation result is used for navigating in game. The novel scheme based on Markov model fusing appearance, color and motion information. The experimental results demonstrate that the proposed approach which is used for 3D game controlling, is real-time and robust, and can give players more immersiveness.

In contrast to the conventional usage of the Hough transform for fingerprint alignment, our novel Hough transform-based fingerprint registration algorithm proposed in this paper doesn't need to discretize the Hough space when estimating the parameters of the rigid transformation between two fingerprint impressions. In our algorithm, the parameters (rotation, x translation and y translation) of the transform (rigid transform considered) can be estimated separately. In addition, the orientation fields from the fingerprints, instead of the minutiae, are used to estimate the most optimal rotation parameter. The experimental results obtained on the public domain collections of fingerprint images, FVC2002 db2 set A (800 fingerprints) and db3 set A (800 fingerprints), show that the new fingerprint registration algorithm presented in this paper is faster and more effective as compared to the conventional generalized Hough transform-based fingerprint alignment method.

Aiming at the segmentation of low quality fingerprint images, a new framework, which is different from traditional methods that usually use some certain features to segment diversified images, is proposed. There are two contributions in this scheme: firstly, we introduce a quality estimation step before segmentation, which can remove a great many false traces effectively; secondly, a new feature eccentric moment is proposed to locate the blurry boundary. Then we segment the image using the new block feature of clarified image. Experimental results show that the proposed method can segment low quality fingerprint images properly.

The most expressive way human display emotion is through facial expressions. Facial expression recognition is necessary for designing any practical human-machine interfaces. This paper proposes a novel framework to real-time facial expression recognition within the interactive computer environment. The two major contributions of this work are: first, we proposed a novel network structure and parameters learning algorithm for embedded HMM (Ara V. Nefian and Monson H. Hayes III, March, 1999) based on AdaBoost (Freund, Y. and Schapire, R. E., 1997). Second, we apply this optimized embedded HMM to real-time facial expression recognition. In this paper, the embedded HMM uses two-dimensional discrete cosine transform (2D-DCT) coefficients as the observation vectors opposite to previous HMM approaches which use pixel intensities to form the observation vectors. Our proposed system reduces the complexity of the training and recognition system. It offers a more flexible framework and can be used in real-time human-machine interactive applications. Experimental results demonstrate that the proposed approach is an effective method to recognize facial expression.

The reliability of facial recognition techniques is often affected by the variation of illumination, such as shadows and illumination direction changes. In this paper, we present a novel framework, called the self-quotient image, for the elimination of the lighting effect in the image. Although this method has a similar invariant form to the quotient image by Shashua etc. (2001), it does not need the alignment and bootstrap images. Our method combines the image processing technique of edge-preserved filtering with the Retinex applications of by Jobson, et al., (1997) and Gross and Brajovie (2003). We have analyzed this algorithm with a 3D imaging model and formulated the conditions where illumination-invariant and -variant properties can be realized, respectively. A fast anisotropic filter is also presented. The experiment results show that our method is effective in removing the effect of illumination for robust face recognition.

Abstract Two approaches, image‐space and object‐space, exist for non‐photorealistic rendering (NPR). The object‐space approach has an advantage that it is able to access the 3D shape information of the scene with a 3D model. In this paper, we present a non‐stroke based image‐space technique that has the strength similar to that of the object‐space approach, but without involving explicit 3D models. The basic idea is to factorise the Lambertian model to obtain the shape and texture information implicitly contained in the image. We also develop a specific anisotropic low‐pass image filter, which can smooth an image without blurring important features. With the developed techniques, we present two NPR styles, the sketching style and the painting style. Although our work does not aim to mimic any specific existing art form, the sketching style can generate effects similar to those of the pen‐and‐ink and the painting style is able to produce the colour permeating effects of watercolour. Copyright © 2004 John Wiley & Sons, Ltd.

Minutiae extraction is one of the most important steps for automatic fingerprint identification systems. However, the performance of minutiae extraction relies heavily on an enhancement algorithm. There are many enhancement algorithms that depend on the local orientation field of the fingerprint. In the singular point area, because the local orientation changes very rapidly, enhancement is not accurate, so that the result is very bad. We present a new method and design a new filter to enhance a fingerprint in the singular point area. We distinguish the singular point area first. Then we design a new filter to enhance this area. Experimental results show a significant improvement of fingerprint enhancement in the singular point area, and the time required for our algorithm is reduced.

We present a new algorithm for recognizing face images under different lighting conditions. Instead of the commonly used methods of modeling the 3D shape or finding some illumination invariant, the basic idea of our algorithm is to create new synthetic images by simulating lighting effects of the input image for each candidate image in the database, and then checking correlation between the input image and the resulting synthesized images. With our method, only a single image is needed for each face in the database, which significantly decreases the size of the database. Theory analysis and experimental results show that this algorithm can strikingly enhance the recognition rate and can be used for real time applications.

Surrounding neighbor blocks dependent fragile watermark scheme has its disadvantages, which couldn't successfully distinguish where the image has been tampered. In this paper, we propose a random block dependant fragile image watermark scheme. Two series, one is generated by this block and its prior block, the other is generated by this block itself, are embedded into the LSBs of this image block to keep itself secure and dependent on its prior block, which make VQ attack impossible. Decision strategy is implemented to detect whether the image blocks are authentic in tamper detection. Analysis and experimental results demonstrate that our algorithm can detect even one bit image alteration with graceful localization, and successfully resist vector quantization attack simultaneously without the need of any unique keys or image indexes.

To deal with image variations due to illumination problem, recently Ramamoorthi and Basri have independently derived a spherical harmonic analysis for the Lambertian reflectance and linear subspace. Their theoretical work provided a new approach for face representation, however both of them had the assumption that the 3D surface normal and albedo are known. This assumption limits this algorithm's application. In this paper, we present a novel method for modeling 3D face shape and albedo from only three images with unknown light directions and this work well fills the blank, which Ramamoorthi and Basri left. By taking the advantage of similar 3D shape of all human faces, the highlight of the new method is that it circumambulates the linear ambiguity by 3D alignment. The experiment results show that our estimated model can be perfectly employed to face recognition and 3D reconstruction.

To deal with image variations due to illumination problem, Ramamoorthi and Basri have independently derived a spherical harmonic analysis for the Lambertian reflectance and linear subspace. Their theoretical work provided a new approach for face representation, however both of them assume that the 3D surface normals and albedo (or unit albedo) are known, which limit this algorithm's application. We present a novel method for modelling 3D face shape and albedo from only three images and this work will fill the blank which Ramamoorthi and Basri left. Our work is closely related to photometric stereo, but conditions of photometric stereo for estimating albedo and surface normal are too strict to be applied for real application. Moreover, the conventionally used singular value decomposition (SVD) approach leads to the notorious linear ambiguity and the solution to this problem needs to introduce more constraints or more images. By taking advantage of similar 3D shape of all human faces, the highlight of the new method is that it circumambulates the linear ambiguity by 3D alignment. The experiment results show that our estimated model can be perfectly employed to face recognition and 3D reconstruction.

Fingerprint classification provides an important fingerprint index and can reduce fingerprint matching time in a large database. A good classification algorithm can give an accurate index that is able to search a fingerprint database more effectively. We present a fingerprint classification algorithm that is based on directional fields. We compute directional fields of fingerprint images and detect singular points (cores). Then, we extract features that we define from fingerprint images. We also use k-means classifier and 3-nearest neighbor to classify features and distinguish which fingerprint is Arch, Left Loop, Right Loop, or Whorl. Experimental results show a significant improvement in fingerprint classification performance. Moreover, the time required for the classification algorithm is reduced.